b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) e 1 – e 3

avai lab le at www.sc iencedi rec t .com

journal homepage: www.e lsev ier .com/ locate /b iochempharm

Correspondence

In response to ‘‘Selective serotonin reuptakeinhibitors—A new modality for the treatment oflymphoma/leukaemia?’’ by C. Schuster, et al.[Biochem. Pharmacol. 74 (9) 2007 1424–1435]

Dear Sirs,

We read with great interest the above paper from Schuster and

colleagues. The study is well performed and we believe has

merit. As the take-home message of the authors is aimed

squarely against our recent and ongoing work we felt it

important to respond by contesting their assertion that: ‘‘SSRIs

are unlikely to represent useful lead compounds for inducing

apoptosis in B-cell derived tumours’’. We take the opportunity

here to present to the readership of Biochemical Pharmacology,

together with the wider scientific and medical community,

additional facts and insight such that an informed judgement to

the question posed by these authors can be reached.

In essence, our own studies (encapsulated in Refs. [1,2])

have shown the following:

(1) E

arly passage, ‘biopsy-like’ Burkitt’s lymphoma cells reveal

a heightened sensitivity to the pro-apoptotic actions of

SSRI antidepressants compared to the constitutive popu-

lations of other (non-BL) lymphoid cell lines; also by

comparison to normal resting B cells.

(2) L

onger-established BL lines and those that have drifted

away from the biopsy phenotype show a relative resistance

to the actions of SSRI antidepressants.

(3) E

xpressing a bcl-2 transgene affords considerable protec-

tion to biopsy-like BL cells from SSRI-promoted apoptosis.

(4) T

he pro-apoptotic outcome can be achieved with the

structurally distinct SSRIs: fluoxetine (Prozac1), paroxetine

(Seroxat1) and citalopram (Celexa1).

(5) R

egards fluoxetine (for which pharmacodynamic and

bioavailability data are most readily available) the con-

centrations at which biopsy-like BL cells are killed in vitro

may be achieved safely in vivo.

DOI of original article: 10.1016/j.bcp.2007.07.017

From the above and associated data we are considering

fluoxetine in a small Phase II clinical trial of Burkitt’s

lymphoma where patients have relapsed, or are refractory,

to the current cytotoxics vincristine/cyclophosphamide/

doxorubicin/methotrexate. The need for new treatment

modalities for this dreadful disease, particularly in

sub-Saharan Africa where BL remains a childhood killer,

was highlighted starkly and eloquently in a recent Lancet

essay [3]. Our approach – i.e. ‘Drug Reprofiling’ – allows

leapfrogging over both animal models and Phase I trials by

virtue of the compound’s safety profile being already

approved.

The reasons we suggest that the conclusion reached by

Schuster and colleagues of dismissing SSRIs for therapeutic

consideration in Burkitt’s lymphoma at this stage is wrong are

as follow:

(I) T

he authors failed to include in their study any Burkitt’s

lymphoma line that could be considered close to the

malignant cells as they exist in vivo. By their very nature

established cell lines are models, at best, for probing

function and behaviour of the original tumour. We have

deliberately concentrated our efforts on studying ‘‘early

passage, biopsy-like’’ BL cells which by all measured

criteria remain phenotypically faithful to their in vivo

precursors. Importantly this includes two of BL’s hallmark

features: extremely high rate proliferation coupled with

high rate spontaneous apoptosis—this seemingly para-

doxical pattern being a consequence of the (Ig locus)

translocated myc gene that defines the disease at a stage

in ontogeny when Bcl-2 is not normally expressed. This

behaviour helps to explain that while BL is one of the

fastest growing tumours it enjoys high cure rates, at least

in industrialised nations. Long-term BL lines progres-

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) e 1 – e 3e2

sively lose the background apoptosis inherent to the

tumour by it being actively selected against during

adaptation to in vitro culture: early passage BL cells retain

this pivotal trait. In Epstein-Barr virus (EBV)-positive BL

artefactual loss of this defining attribute can develop

within a year due to in vitro liberation of the otherwise

suppressed transforming EBV-latent gene programme:

this is described as a shift from Latency (LAT) I to Latency

III. The three BL lines studied by Schuster and colleagues

were first established in the 1960s and two (Namalwa and

Daudi) harbour EBV. Where we included one of these

(Namalwa) in comparative analysis for a growth arrest

response to fluoxetine it was noticeably less sensitive

than early passage BL lines while isogenic sublcones of

the EBV+ve BL line Mutu were appreciably more sensitive

in their natural LAT I state compared to in vitro-evolved

Lat III variants [1].

(II) F

or reasons already mentioned, our proposed ‘lead

compound’ for potential translation of SSRIs to BL

therapy is fluoxetine [1,2]. Surprisingly, Schuster and

colleagues omitted fluoxetine from their study. Instead,

they focused on paroxetine and citalopram, which does

allow us some comparison. As can be noted from our

original report [1], against early-passage L3055 BL cells we

obtained IC50 values (anti-proliferative) for each of these

two SSRI lower than those observed against any of the

long-established cell lines investigated by Schuster et al.:

6.9 � 1.2 mM with paroxetine; 20.9 � 1.8 mM with citalo-

pram. Interestingly – and perhaps particularly telling –

the single most sensitive line in Schuster and colleagues’

work is that generated from murine Em-myc tumour cells:

as with the early passage BL cells we work with, they

seemed to study these within months of isolating ex vivo

as opposed to decades for the BL lines used. The Em-myc

cells revealed a level of sensitivity (IC50 = 5.5 � 1.46 mM)

to paroxetine nigh identical to that found by us against

biopsy-like BL cells [1]. Such agreement between the two

studies stimulates confidence in the value of the

approach we are suggesting.

(III) W

e feel persuaded that the heightened sensitivity of

(biopsy-like) BL cells to SSRIs (and to pro-apoptotic

modalities generally) is due in part to their negligible

expression of anti-apoptotic Bcl-2 as evidenced by the

protection that is afforded once it is ectopically expressed

[1,2]. However, among different cell lines of assorted

pathologies additional, even alternative, survival path-

ways will come into play: as we have indicated ourselves

in a separate study with regards sensitivity among a

diverse panel of B-cell lines to the pro-apoptotic actions of

dopaminergic drugs [4]. Also, even among biopsy-like BL

cells (including L3055) survival can be contributed by

further Bcl-2 family members and unrelated anti-apop-

totic genes such as cIAP2 and A20 [5]. It is therefore

unsurprising that Schuster and colleagues found no

simple correlation between Bcl-2 levels and sensitivity

to paroxetine and citalopram among the broad range of

lines they studied. Furthermore, while Bcl-2 strongly

protects from the pro-apoptotic actions of antidepres-

sants it affords minimal protection to their anti-prolif-

erative outcomes [2].

(IV) W

ith regards therapeutic considerations, apoptosis is

undoubtedly the optimal desired outcome. Unfortunately

the single measure used by Schuster et al. to gauge

‘‘apoptosis’’ is Annexin V-binding. Though we have used

this when requested by reviewers [1] we prefer to avoid

this method and certainly would never employ it as a sole

indicator of apoptotic death. As exemplified in Ref. [6]

phosphatidylserine exposure (which is what AV-binding

detects) is far from restricted to apoptotic outcomes and

can accompany multiple activation pathways in a

diversity of cells. It is therefore difficult to assess bona

fide apoptosis from the paper of Schuster and colleagues.

Again we would like to emphasise that despite their main

conclusion being an inaccurate and misleading extrapolation,

there are elements of Schuster and colleagues’ study that

make a valuable contribution to this new and exciting field.

Particularly pertinent is the demonstration that the major

common target for which the SSRIs were first developed as

antidepressants – i.e. the serotonin transporter, SERT – is

unlikely to be responsible for the anti-lymphoma actions

being discussed here. Though our initial rationale for pursuing

SSRIs as potential BL therapeutics arose from having identified

SERT expression in BL cells, subsequent study has provided us

circumstantial evidence leading us to concur that SERT is not

the target by which SSRIs exert anti-BL activity [1,2].

Whatever the functional target might be, we feel for the

reasons discussed above that it is premature to dismiss SSRIs

for consideration as potential treatment adjuncts in Burkitt’s

lymphoma. We have reviewed the available literature regard-

ing concentrations of fluoxetine that can be reached in vivo on

current dosing and discussed how these translate to what is

observed and necessary for BL cell killing in vitro [1,2]. At the

current maximal on-license dosing of 80 mg/day (for OCD,

obsessive compulsive disorder) the concentrations required

for BL kill may be achievable. With these fluoxetine concen-

trations, neither peripheral blood mononuclear cells nor

normal purified B cells succumb to apoptosis; as neither do

long-established cell lines, bcl-2-carrying BL, nor those that

have drifted to an in vitro LAT III phenotype. The conventional

cytotoxics currently used in treatment can by no means be

considered ‘‘specific’’; even a targeted, rationally designed

drug such as Rituximab hits all B cells. The observed

heightened pro-apoptotic sensitivity of (genuine) Burkitt’s

lymphoma cells to fluoxetine indicates a rare ‘Therapeutic

Window of Opportunity’ to be tried and tested in individuals

that are still dying from this most terrible disease.

Acknowledgment

Work of the above is supported by Leukaemia Research (U.K.).

r e f e r e n c e s

[1] Serafeim A, Holder MJ, Grafton G, Chamba A, Drayson MT,Luong QT, et al. Selective serotonin reuptake inhibitorsdirectly signal for apoptosis in biopsylike Burkitt lymphomacells. Blood 2003;101:3212–9.

b i o c h e m i c a l p h a r m a c o l o g y 7 5 ( 2 0 0 8 ) e 1 – e 3 e3

[2] Meredith L, Holder MJ, Chamba A, Challa A, Drake Lee A,Bunce CM, et al. The serotonin transporter (SLC6A4) ispresent in B-cell clones of diverse malignant origin: probinga potential anti-tumor target for psychotropics. FASEB J2005;19:1187–9.

[3] Phillips JA. Is Burkitt’s lymphoma sexy enough? Lancet2006;368:2251–2.

[4] Meredith L, Holder MJ, Rosen A, Drake Lee A, Dyer MJ, BarnesNM, et al. Dopamine targets cycling B cells independent ofreceptors/transporter for oxidative attack: implications fornon-Hodgkin’s lymphoma. Proc Natl Acad Sci USA2006;103:13485–90.

[5] Stewart R, Wei W, Challa A, Armitage RJ, Arrand JR, Rowe M,et al. CD154 tone sets the signaling pathways andtranscriptome generated in model CD40-pluricompetentL3055 Burkitt’s lymphoma cells. J Immunol (in press).

[6] Holder M, Barnes NM, Gregory CD, Gordon J. Lymphomacells protected from apoptosis by dysregulated bcl-2continue to bind Annexin V in response to B cell receptorengagement: a cautionary tale. Leukemia Res 2006;30:77–80.

John Gordon*

Nicholas Barnes

Elisabeth Meredith

Mark Drayson

Division of Immunity & Infection,

The Medical School,

Birmingham B15 2TT, UK

*Corresponding author at:

The University of Birmingham,

MRC Centre for Immune Regulation,

Vincent Drive, Birmingham B15 2TT, UK.

Tel.: +44 121 414 4034; fax: +44 121 414 3599

E-mail address: j.gordon@bham.ac.uk (J. Gordon)

0006-2952/$ – see front matter# 2008 Published by Elsevier Inc.doi:10.1016/j.bcp.2008.02.009